While user journal replication has significant advantages, it may not be appropriate for all of the files in your environment. Or, it may be appropriate for only some of the supported object types.
User journal replication for journaled IFS objects, data areas, and data queues may be required or beneficial for the following:
-
User journal replication is recommended if any of the following are true:
-
Small modifications are made to files by adding or changing a subset of the file
Note: An application like Microsoft Word or Microsoft Excel will typically re-write theentire file each time the file is saved, so minor end-user edits to these files do not
fall into this category.
-
Files are moved, renamed, deleted, and/or recreated rapidly.
-
-
If multiple hard links to a single stream file are used, user journal replication is required.
System journal replication for journaled IFS objects, data areas, and data queues may be required or beneficial for the following:
-
IFS objects that meet the following criteria are better suited for replication through the system journal:
-
Are written or re-written entirely at one time
-
Are rarely modified for changes or additions to only a small portion of their data
-
Do not perform many rapid move/rename/create/delete operations involving the same names
For example, applications that store scanned images for archival purpose and repositories for office documents typically meet this criteria.
-
If you want to omit data from *DTAARA or *DTAQ objects so that only the “shell” of an object is replicated, system journal replication is required. This can be configured so that only these object types or specific objects within your configuration will be replicated through the system journal.
The benefits of user journal replication are described in User journal replication of IFS objects, data areas, data queues. For restrictions and limitations, see Identifying data areas and data queues for replication and Identifying IFS objects for replication.